3.1 Definition, Basel event types, consequential effects
▼The Basel definition syllabus 3.1.1
The definition covers legal risk (fines, penalties and punitive damages resulting from regulatory actions, and private settlements) but excludes reputation risk. The four causes to remember: processes, people, systems, external events.
The workbook notes the techniques to manage it are very diverse because there is little commonality between people, processes, IT systems and external events (e.g. bomb threats, power cuts).
The seven Basel event types syllabus 3.1.2 - 3.1.3
| Event type | Loss arising from... |
|---|---|
| Internal fraud | Acts intended to defraud, misappropriate property or circumvent regulations, law or company policy (excluding diversity/discrimination events), involving at least one internal party |
| External fraud | Acts intended to defraud, misappropriate property or circumvent the law by a third party |
| Employment practices and workplace safety | Acts inconsistent with employment, health or safety laws or agreements, personal injury claims, or diversity and discrimination events |
| Clients, products and business practices | An unintentional or negligent failure to meet a professional obligation to specific clients (incl. fiduciary and suitability requirements), or from the nature or design of a product |
| Damage to physical assets | Loss or damage to physical assets from natural disaster or other events |
| Business disruption and systems failures | Disruption of business or system failures |
| Execution, delivery and process management | Failed transaction processing or process management, or relations with trade counterparties and vendors |
The BIS says that for banks of any size the crucial elements of an effective operational risk framework are clear oversight by the board and senior management, a strong operational risk culture, and a strong internal control culture (clear lines of responsibility, segregation of duties, effective internal reporting, contingency planning). The workbook also notes loss types evolve and the industry consortium ORX has published a new reference taxonomy for operational and non-financial risk.
A distinct risk class, and its knock-on effects syllabus 3.1.4
Market and credit risk are relatively easy to set specific measurable levels for; operational risk is hard to identify or assess numerically. Historically firms accepted it as an unavoidable cost and treated anything that was not market or credit risk as operational. Modern firms have failed through operational causes too: malfunctioning processes, technology failure, undesirable staff behaviour, fraud, compliance issues.
When an operational risk materialises it often triggers other risks:
| Consequence | How it arises |
|---|---|
| Reputational | Clients or media learn of the issue, leading to lost business and revenue |
| Compliance / regulatory | Process failures, e.g. customers not treated fairly, breach rules and can bring fines or sanctions |
| Credit | Data errors causing inadvertent credit limit breaches; poor monitoring or misreading of counterparty accounts through lack of training; legal risk such as being unable to enforce contracts (e.g. collateral posting); skipping credit checks or assuming rating agencies are always right |
| Market | An undetected error in the portfolio management system causes a market risk limit breach |
| Liquidity | A finance department process breakdown leaves too little liquidity to pay salaries |
| Investment | A careless fund manager plus no checking process causes a mandate limit breach |
3.2 Operational risk policy
▼What the policy is for syllabus 3.2.1
A firm needs a written operational risk policy giving a coherent, consistent approach: a "roadmap" from a fragmented approach to a firm-wide methodology with a common risk language. It outlines strategy and objectives, and clarifies the boundary between operational and other risks such as market and credit. Building the framework involves:
- defining operational risk appetite
- defining how risks are identified and categorised
- defining how risks are measured and assessed for significance
- assigning line managers ownership of mitigating actions to bring exposure within appetite
- assigning responsibility for monitoring the effect of those actions
- setting reporting and escalation mechanisms for transparency and decision-making
Developing the policy is cyclical and continuous. A common global policy and terminology lets the firm balance global standardisation against regional differences, prioritise risks objectively, reward or penalise risk performance fairly, and centralise the capital adequacy assessment. Because it cuts across departments, a central risk management role (possibly an independent department) coordinates it.
The four areas the policy addresses syllabus 3.2.1
| Area | Key points |
|---|---|
| 1. Identification of key officers | Line managers in the independent operational risk function (monitor and report to the board); senior business managers (risks in their area); the group risk management function; and risk representatives / risk champions in the business, often with dual reporting lines to their business area and to risk management |
| 2. Roles and responsibilities | Clear responsibility and accountability at all levels; people accountable need the control and authority to act; clear lines of authority; consequences of non-compliance made clear |
| 3. Segregation of duties | Separate front office, operations, accounting and risk monitoring. Barings: the head of front office in Singapore also headed the back office and could hide trading errors until they sank the bank. Since then some regulators bar mixing roles such as head of compliance and sales |
| 4. Cross-functional involvement and agreement | Encourage collaboration because many risks arise from ownerless or unnoticed boundary errors, without giving up segregation of duties, through reporting, incentives, education and supportive structure |
3.3 The framework: aims and stages
▼Key aims of the operational risk function syllabus 3.3.1
Many risks are best managed in the departments where they arise (IT staff for systems risks, back office for settlement). But overall planning, challenge and monitoring come from a central operational risk function independent of the business. Its role includes:
- working with risk owners to assess and quantify risks
- providing a dotted or solid reporting line for business risk representatives
- maintaining the operational risk system that tracks risks, controls, losses, actions and key indicators
- benchmarking good industry practice; oversight and monitoring
- ensuring issues are escalated and tracking actions from incidents
- qualitative analysis: loss causal analysis, HR exit-interview reports, internal audit reports and how quickly audit points are closed
- statistical modelling to quantify the risk profile for regulatory and other purposes
The three aims: identify and assess risks, manage them, and reduce potential impact and likelihood.
Common identification methods: self-assessment, KRIs, risk and control assessment workshops, loss data causal trend analysis, external loss data (where available) and audit reviews, all needing the full involvement of risk owners.
Reducing likelihood vs reducing impact syllabus 3.3.1
| Goal | How |
|---|---|
| Reduce likelihood | Identify the risk before it occurs; establish clear ownership and let the owner put controls in place; set and monitor risk indicators and act before they reach predefined danger limits |
| Reduce impact (if it happens) | Speedy escalation to senior management; an owner assigned to fix the problem who can draw in other resources; appropriate insurance if necessary |
The framework stages syllabus 3.3.2
The framework starts and ends with the operational risk policy and appetite setting, so lessons learned feed back into the policy:
| Stage | What happens |
|---|---|
| Risk identification | Clearly identify risks using the methods above |
| Measurement and assessment | Score impact and likelihood against pre-defined criteria |
| Management and control | Ensure appropriate controls; action plans for under-controlled risks; real-time escalation with pre-defined thresholds saying how high breaches or losses must be escalated; remediation owned and tracked to completion |
| Risk monitoring | Monitor risk and control indicators and other management information (MI); act before predefined danger limits are reached |
| Risk reporting | Report indicators, related risks and controls, and incidents, ideally losses AND near misses. Triggers = amber (management attention); limits = red (immediate action) |
| Policy | Update the policy from lessons learned |
3.4 Identification and categorisation
▼Why categorise risks syllabus 3.4.1
Identifying and categorising risks helps a firm establish its risk profile and appetite. A natural starting set of categories is the definition itself: process, people, system and external event risks, subdivided further. Categorising allows:
- more succinct MI grouped by category instead of a long list of individual risks
- a better understanding of where weakness lies (processes, systems, people, external events)
- a sound basis for allocating operational risk capital across categories
- a common language for discussing and managing risk
Self-assessment (self-certification) syllabus 3.4.2
Typically a checklist of the risks an area faces; managers and staff score each risk (survey or questionnaire), usually by probability and impact. Limits as a single method:
- Subjective and open to abuse or manipulation by managers, so it should be independently validated, e.g. by internal audit or a risk officer who does not report to the department head.
- Combining scores into one score per risk is difficult: unclear whether to average or to investigate outliers.
A common fix is a workshop with key staff to brainstorm risks to departmental objectives, debating until a consensus emerges on impacts and likelihoods.
Applying the four causes to examples syllabus 3.4.3
| People | Process | System | External event |
|---|---|---|---|
| Inadequately defined roles and responsibilities | Lack of written procedures | System unavailable during peak hours | Threat of terrorist action |
| No succession plan for key staff | Absence of a defined process | Data becomes subtly corrupted | Customer commits fraud against the firm |
| Staff not competent for role | Manual intervention causes pinch-points | Passwords shared by staff | Outsource supplier delivers late |
| Improper staff conduct | Absence of escalation procedures | Denial of service (DoS) attacks | Inability to access key premises through disaster |
3.5 Assessment: impact x likelihood, scenarios, bottom-up
▼Why measure, and why it is hard syllabus 3.5.1 - 3.5.2
Assessment and measurement both aim to understand the likelihood of risks and their potential impact. Reasons to do it:
- establish a quantitative baseline for improving controls (e.g. what avoiding the risk might save, for the business case)
- provide an incentive for risk management and a strong risk culture
- improve management decisions on how much risk to take
- satisfy regulators and shareholders that the firm is proactive and transparent
- assess financial exposure for capital allocation
Main difficulty: a lack of relevant, objective data; many firms lack enough own historic loss data. Banking and insurance industries have tried to solve this by anonymously sharing losses.
| Term | Meaning |
|---|---|
| Risk measurement | Quantitative techniques for the size of the risk profile, e.g. statistically modelling frequency and impact and predicting the future profile |
| Risk assessment | Uses whatever objective data exists (e.g. the measurement output) plus human judgment to estimate business impact; with no data, subjective experience alone |
Example: risk indicators (a measurement) show the front office trading system is 98% reliable; a risk assessment then judges whether that is acceptable.
Impact and likelihood assessment syllabus 3.5.3
One of the simplest methods: rank risks by severity using two criteria, likelihood and impact. It may be subjective, objective (supported by history) or both. The workbook gives illustrative rating scales (the actual bands are firm-specific):
| Rating | Likelihood (example) | Impact (example) | Score |
|---|---|---|---|
| Very low | Not likely within the next ten years | Under £1,000 | 1 |
| Low | Likely within three to ten years | £1,000 to £10,000 | 2 |
| Medium | Likely within two to three years | £10,000 to £50,000 | 3 |
| High | Likely within the next year | Above £50,000 | 4 |
There is no "correct" number of bands, but four each is common, and an even number is preferred so risks are not scored "medium" (the middle) by default. Scores rank risks and can be plotted on a heat map coloured red, amber, yellow and green to show which risks are inside or outside tolerance. The workbook's example map has green in the low-likelihood/low-impact corner and red in the high-likelihood/high-impact area.
Firms often score gross (assume no controls / all controls failed) and net (after considering the control environment), which lets the effectiveness of controls be evaluated separately from the risks they mitigate.
- 1 Likelihood: "likely within the next year" = High = 4.
- 2 Impact: £30,000 sits in the £10,000 to £50,000 band = Medium = 3.
- 3 Risk score = 4 × 3 = 12. A different risk with likelihood 2 and impact 4 scores 8, so it would rank lower even though its impact rating is higher.
Impact and likelihood: strengths and weakness syllabus 3.5.3
| Advantages | Disadvantage |
|---|---|
| Simple view of the range of risks; shows control effectiveness if gross and net are plotted; focuses attention on the most important risks; works with minimal hard data; captures large strategic risks and everyday issues at all levels; encourages a risk-aware, transparent culture through continuous assessment | May be over-simplified and subjective, so validate with real loss data and/or an independent party (internal audit, central risk function, peer review) |
Scenario analysis vs bottom-up analysis syllabus 3.5.3
| Scenario analysis | Bottom-up analysis | |
|---|---|---|
| Direction | Top-down | Bottom-up: builds a detailed profile of risks in each area and aggregates to departments, divisions and the firm |
| Idea | Uses business professionals' experience to capture scenarios that occurred or may cause loss, highlighting risk combinations; may stress (exaggerate) an aspect so preventative measures can be taken | Analyses individual risks and control adequacy across business processes, using line managers' and staff experience plus loss data (qualitative and quantitative) |
| Advantages | Enables preventative action against combinations of events | Works at process level; clear accountability (process owner is accountable); transparent, risk-aware culture; continuous improvement; better MI |
| Disadvantages | - | Slow to implement (documenting processes across departments); can be subjectively influenced by managers; aggregating upwards is not straightforward (assigning risks to categories helps) |
3.6 KRIs, historical loss data, practical constraints
▼Key risk indicators (KRIs) syllabus 3.5.4
After ranking risks, a firm designates its top "x" risks as key risks, obtains data describing their current status, and sets upper and lower acceptable limits: these produce KRIs. Example for "loss of key staff": indicators such as general staff turnover (as a proxy), percentage of undocumented processes, and salary gaps found by annual benchmarking. Some firms use a "soft" amber limit to act before the "hard" red limit; adjusting amber and red limits over time also helps set risk tolerance.
| Advantages | Disadvantage |
|---|---|
| Monitor trends and anticipate problems; establish limits of acceptability; provide a basis for objective measurement | Managers may "manage to their KRIs" to boost bonus ratings, skewing performance |
Key design point: build indicators on the drivers of likelihood of a key risk. They are then forward-looking and predictive and flag increased risk before losses occur. The trade-confirmation case study gives KRIs such as time taken to formalise a legal agreement, number of confirmations not yet agreed, number and type of confirmation errors found, and time taken for counterparties to return confirmations.
Historical loss data and the loss distribution syllabus 3.5.5
Losses (internal or external) are mapped to a sensible categorisation and used with benchmarking or statistical methods. A loss distribution curve records the value of material direct losses in a category over a period (e.g. three years); VaR-type techniques allow predictions within confidence limits. The curve has a fat tail because losses are not normally distributed: extremely high-impact losses are very rare.
| Loss type | Frequency / cost | How it is paid for / managed |
|---|---|---|
| Expected | Low cost, high frequency | Pre-provisioned in the annual budget or paid from current revenues; within risk appetite or tolerance; known weaknesses that are perhaps too expensive to fix |
| Expected | Low cost, low frequency | Paid from current revenues |
| Unexpected | High cost, low frequency | Covered by operational risk capital or insurance; hard to manage day to day because they rarely test the controls; managed with contingency planning; cost borne by the risk capital buffer |
Practical constraints syllabus 3.5.6
| Constraint | Issue |
|---|---|
| Data collection | Hard to build a comprehensive data set: general lack of data, system constraints, lack of standardisation, little industry-wide data (relies on self-reporting; high-impact low-frequency events are hard to understand); some firms may be barred from reporting for legal reasons |
| Cultural | Business heads need convincing; ORM can be seen as a cost or nuisance, so many firms roll out little by little, winning one area's support before moving on |
| Resource and cost | Firms underestimate the time and resources needed to implement identification and measurement systems |
| Indicator | Hard to design indicators covering the full range of risks; tendency to reuse existing MI, which is often built to monitor performance rather than risk |
3.7 Risk register + the four response methods
▼The risk register (risk log) syllabus 3.6.1
After identifying, assessing and ranking risks, a firm builds a risk register (also called a risk log). There is no widely accepted format; the workbook's illustrative core features are:
- objectives, processes or products affected by the risk
- description of the risk
- risk ranking (the overall score is used to sort the list from most to least significant)
- lead person or department (owner)
- action plan, with target and completion dates
- sources of assurance and oversight (may or may not be the lead person or department): mechanisms giving an objective view on the quality of each risk's management
- mitigating controls, their effectiveness and owners
- gross, net and residual risks
Mitigating controls reduce likelihood and impact, e.g. managerial sign-off before work leaves a department, or an IT system requiring a second person to verify inputs.
Transfer, avoid, mitigate, accept syllabus 3.6.2
| Response | Meaning | Examples in the workbook | Controls needed? |
|---|---|---|---|
| Transfer | Contractually pass the risk to a third party better equipped to deal with it | Outsourcing; insurance | No (for the transferred part) |
| Avoid | Used when the level is unacceptable and managing it would be too costly | Withdrawing from a business or a product with unmanageable errors; changing a product offering; not taking on new business through planned mergers or acquisitions | No |
| Mitigate | Reduce likelihood and/or impact; the most common action is introducing key controls into processes | Controls, training, systems | Yes |
| Accept | Not spending resources to mitigate, simply accepting the possible consequences | "Risk acceptance" | No |
The workbook stresses that controls are only required when the business has accepted (retained) the risk in the sense of mitigating it: neither avoiding nor transferring it needs controls.
3.8 Controls, financial crime, resilience
▼Preventative vs detective controls syllabus 3.6.3
All processes should have controls and check points to detect errors and prevent fraud and theft. Example: segregation of duties means no one person has end-to-end authority (one books a payment, another checks it), preventing e.g. cash being assigned to personal accounts.
| Type | When it works | Examples |
|---|---|---|
| Preventative | Stops errors occurring in the first place; tackles root causes; most effective if built into processes at the outset; technology often used | Individual IT passwords and access control; procedures preventing unauthorised actions; training to reduce human error; automating processes |
| Detective | Detects errors once they have occurred | Quality assurance checks |
Trade confirmation controls (worked case): a legal agreement on confirmation protocol before trading; front-office sign-off of each trade's economic terms; a back-office confirmation check; daily follow-up with counterparties who have not returned written confirmations.
Processes change, so the control structure must be reviewed as part of change; identifying control gaps is a key objective of the operational risk function.
Financial crime compliance syllabus 3.6.3
Financial crime prevention (fraud, bribery and corruption, cybercrime, money laundering, terrorist financing) is treated as one of the common methods of operational risk management. Money laundering turns "dirty" money into money that appears to have legitimate origins, because dirty money is hard to invest or spend and can be evidence of the original crime.
| Stage | Meaning |
|---|---|
| 1. Placement | Introducing dirty money into the financial system, e.g. cash into a bank or building society account, bureau de change, or any business routinely accepting large amounts of cash |
| 2. Layering | Moving placed money around to hide the link to the beneficiary: rapid currency, share or bond trades, collective investment schemes, insurance-based investment products, moving money between countries |
| 3. Integration | Layering has succeeded and the beneficiary appears to hold legitimate (clean) funds |
AML rules broadly require firms to: identify customers and report suspicious transactions at placement and layering; keep adequate records (which should prevent integration being reached); and report suspicious activity to the relevant authority. Terrorist financing (financial support to terrorists or groups) has similar requirements: customer identification, record-keeping and suspicious activity reporting. A person generally commits an offence by entering into or being linked with an arrangement that facilitates retention or control of terrorist funds.
Firm responses: educate staff on the risks (to society, to the firm from regulatory censure, and to the individual from custody or fines); put systems and controls in place; monitor staff compliance; escalate behavioural exceptions to an individual or committee; penalise contraventions and, if necessary, inform the authorities.
Operational resilience: BCP, DR, crisis management syllabus 3.6.3
| Plan | Covers |
|---|---|
| Business continuity plan (BCP) | Premises and people: where staff work if the main site is out of action |
| Disaster recovery (DR) | IT and other infrastructure needed to keep the business running |
Building robust BCP and DR needs analysis of causes of disruption from minor mishaps to catastrophes; risks that could require temporary alternative accommodation include fire, power failure, infrastructure failure, civil unrest and strikes, and terrorism. Both must be regularly tested and shortcomings fixed. Firms also define crisis management teams (CMTs) and methods for crisis management and business resumption; CMTs should rehearse disasters periodically.
3.9 Outsourcing, insurance, cyber, data, physical, training, loss data
▼Outsourcing and insurance (transfer methods) syllabus 3.6.3
| Method | Watch-outs |
|---|---|
| Outsourcing | Passes an activity to a third party with specific expertise. But the firm only transforms the risk: from managing a process directly to managing the quality of the outsourced process |
| Insurance | Can cover fire, theft and human error losses. The firm must be very clear what it pays out for and when. Litigation cover that pays only after judgment may come too late if the firm has already suffered substantial losses |
Information and cyber security syllabus 3.6.3
Information can be on paper as well as in IT systems. Categorise information (personal staff or customer data needs more care than an already-public report) and give staff guidelines (e.g. "personal" information may only be emailed if password protected). Lock important information away at night; many firms have a clear desk policy, which guards against theft and against loss if a building is damaged (bombs blowing documents onto the street) or water pipes burst. Paper-only important records go in a fire-proof safe.
The UK government's "10 Steps To Cyber Security" guidance, as summarised in the workbook:
| Step | Idea |
|---|---|
| Information risk management regime | Assess risks to information assets as vigorously as any other risk |
| Secure IT systems | Remove or disable unnecessary functionality; keep systems patched (vendor updates closing discovered weaknesses) |
| Network security | Filter traffic at the perimeter, monitor for unusual activity, test filters with regular penetration tests |
| Penetration testing | Also "pen testing" or "ethical hacking": systematically probing networks and applications for vulnerabilities |
| Managing user privileges | Users get only the privileges they need; limit privileged accounts (system or database administrators) and keep them off day-to-day activity |
| User education and awareness | User security policies acknowledged in employment terms; regular training; specialist training for security roles |
| Incident management | Incident response and disaster recovery capability, regularly tested; report online crimes to law enforcement |
| Malware prevention | Policies for email, web browsing, removable media and personal devices; antivirus scanning |
| Monitoring | Continuously monitor inbound and outbound network traffic for unusual activity |
| Removable media controls | Limit media types, users and data; scan all media before import |
| Home and mobile working | Assess risks, train mobile users, encrypt data if the device supports it |
The workbook lists these under the "10 Steps" heading, though eleven items appear in its text. You do not need to memorise a count; know the themes.
Data protection and GDPR syllabus 3.6.3
- Under the EU GDPR, fines of up to €20 million or 4% of global revenue, whichever is higher, can be levied on any company with European customers that mishandles their data.
- Customers must be able to give and retract consent, including the right to be "forgotten" (all data deleted). Firms must control and track data internally and across inter-firm boundaries.
- Organisations processing personal data on a large scale need an independent, adequately qualified data protection officer.
- Similar laws named in the workbook: Turkey (24 March 2016), California Consumer Privacy Act (effective 2020), Virginia (2 March 2021), Colorado (8 July 2021), China's PIPL (in force 2021), Saudi Arabia's Personal Data Protection Law (effective March 2022; penalties include up to two years' imprisonment).
These dates are as given in the Edition 11 workbook.
Physical security, financial reserves, risk awareness training syllabus 3.6.3
| Method | Points |
|---|---|
| Physical security | Vet all staff and contractors for criminal records; visible ID cards; sign-in for all visitors; stay vigilant for external threats such as protests or marches aimed at financial firms |
| Financial reserves | Regulatory capital and liquidity requirements (the workbook points back to the regulation chapter) |
| Risk awareness training | Helps staff understand how reducing the likelihood of risk works and their role in it; details and attendance recorded and tracked by the operational risk function |
Using loss data to manage risk syllabus 3.6.4
| Use | How it works |
|---|---|
| Escalation thresholds | Losses of various sizes escalate to pre-defined levels so governance bodies can monitor losses firm-wide and direct resources to areas outside appetite or tolerance |
| Loss causal analysis | Understand root causes so similar losses can be prevented elsewhere. Fixing only the visible effect means the incident will probably recur. Needs pre-defined standard causal categories, otherwise each loss gets a differently worded cause and trend analysis is very difficult. Many firms find the Basel event categories an insufficient taxonomy and extend them with their own causal categories |
3.10 Ch 3 cheat sheet
▼All the numbers + names
| Item | Answer |
|---|---|
| Definition | Loss from inadequate or failed internal processes, people, systems, or external events. Includes legal risk; excludes reputation risk |
| 7 Basel event types | Internal fraud · External fraud · Employment practices and workplace safety · Clients, products and business practices · Damage to physical assets · Business disruption and systems failures · Execution, delivery and process management |
| Policy: 4 areas | Key officers · Roles and responsibilities · Segregation of duties · Cross-functional involvement |
| Barings | Segregation of duties failure (front and back office run by one person) |
| Framework stages | Policy/appetite → identification → measurement/assessment → management and control → monitoring → reporting → back to policy |
| Trigger vs limit | Trigger = amber; limit = red |
| 4 causes | People · Process · System · External event |
| Risk score | Likelihood × impact (example bands 1-4, even number of bands) |
| Gross vs net | No controls vs after controls |
| Scenario analysis | Top-down, risk combinations |
| Bottom-up | Process-level risks and controls aggregated upwards |
| KRIs | Built on likelihood drivers; forward-looking; amber (soft) and red (hard) limits |
| Expected losses | Budget / current revenues; within appetite |
| Unexpected losses | High cost, low frequency; operational risk capital or insurance |
| Constraints | Data · Cultural · Resource and cost · Indicator |
| 4 responses | Transfer · Avoid · Mitigate · Accept |
| Control types | Preventative (stop) · Detective (find, e.g. QA) |
| Money laundering | Placement → Layering → Integration |
| BCP vs DR | Premises and people vs IT and infrastructure |
| Outsourcing | Transforms the risk into managing the quality of the outsourced process |
| GDPR | Up to €20m or 4% of global revenue, whichever is higher; DPO for large-scale processing |
| Loss data uses | Escalation thresholds · loss causal analysis (standard causal categories) |